Assembly and entry mechanisms of Semliki Forest virus

Abstract: The genomes of enveloped viruses are packaged into a core or nucleocapsid (NC) which is surrounded by a lipid membrane with spike proteins. These viruses mature by budding at a membrane of the host cell, usually the plasma membrane (PM), and penetrate into new ones by a virus-host membrane fusion event. The membrane fusion process is, in the case of all enveloped viruses, directed by the spike proteins. In many viruses the spikes are also directing the budding process. The subsequent expression of virus budding and entry processes during virus replication requires that the corresponding functions of the spikes are tightly controlled. I have in this work studied how the functions of the spikes of an alphavirus, Semliki Forest Virus (SFV), are controlled. The SFV spike is composed of three E2-E1 heterodimers. E1 and E2 are two different transmembrane glycoproteins. The NC consists of the viral RNA genome in complex with one kind of capsid (C) proteins. All structural proteins are synthesized from a common coding unit in the order C-p62-E1. The individual proteins are generated through cotranslational cleavages mediated by the capsid protein and the host signal peptidase. Virus budding is directed by NC-spike (E2) interactions and virus fusion by E1-cell membrane interactions. The fusion activity of newly synthesized E1 is suppressed by its association with the E2 precursor protein p62. Activation occurs in two steps. These are (i) cleavage of p62 to E2 in the infected cell and (ii) dissociation of the E2-E1 heterodimer in the virus spike by low pH in endosomes of cells to be infected. In the present study I have investigated the synthesis of the p62-E1 heterodimer with the suppressed fusion function. Using various SFV deletion variants, which were engineered in vitro, I found that the E1 fusion protein never entered a free pool in the endoplasmic reticulum (ER), but it was complexed and chaperoned by the p62 protein immediately after its synthesis. My results suggested that complex formation took place between p62 and E1 proteins originating from the same polyprotein translation product. Most likely, the p62 protein, which is made before E1, waits in the ER translocon until the E1 part has been completed and then associates with it. According to my model the p62 protein can in this way chaperone an intermediate folding stage of E1 which is still inactive in its membrane fusion function. If E1 is synthesized in the absence of p62 it mostly aggregates into a non functional form in the ER. In contrast, p62 can be synthesized without E1 into a form that is transported to the cell surface and also cleaved to E2. However, this form of E2 cannot bind to NCs. This suggests that also the activity of p62/E2 is controlled by E1. Furthermore, I have also in this work addressed questions about the SFV penetration mechanism. In one piece of work I have used in vitro mutagenesis to investigate the role of the short internal protein domain of E1 (--Arg-Arg-COOH) for its membrane fusion activity. I found no correlation between its structure and the fusion activity. In another piece of work I have used various SFV deletion variants to find out which structural proteins of the virus are involved in ion channel formation in infected cell membranes. Such channels are induced by low pH and it has been speculated that they might function in the viral envelope as proton channels during virus entry. I found that the E1 protein was mandatory for the channel activity. Key words: Semliki Forest virus, virus assembly, protein oligomerization, virus budding ISBN 91-628-2151-2